Rheological Characterization of Polymeric Frustrated Lewis Pair

3 days ago - The frequency range for those experiments was from 0.1 to 100 rad s–1. .... Table 2. Details of the Polymer Networks Prepared by the Ad...
0 downloads 0 Views 2MB Size
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Rheological Characterization of Polymeric Frustrated Lewis Pair Networks Utku Yolsal,†,‡ Meng Wang,† John R. Royer,§ and Michael P. Shaver*,†,‡ †

School of Chemistry, University of Edinburgh, Joseph Black Building, David Brewster Road, Edinburgh EH9 3FJ, U.K. School of Materials, University of Manchester, Oxford Road, Manchester M13 9PL, U.K. § School of Physics and Astronomy, University of Edinburgh, King’s Buildings, Peter Guthrie Tait Road, Edinburgh EH9 3FD, U.K. ‡

Macromolecules Downloaded from pubs.acs.org by IDAHO STATE UNIV on 04/26/19. For personal use only.

S Supporting Information *

ABSTRACT: Conventional Lewis acids and Lewis bases will react with each other to form thermodynamically stable Lewis adducts which are inherently unreactive. Sterically encumbered Lewis acids and Lewis bases are prevented from forming such adducts, resulting in the formation of frustrated Lewis pairs (FLPs) with latent reactivity. Extending this concept to polymer science permits the development of a new class of responsive, functional, self-healing materials. Here, we report the rheology of fully macromolecular FLPs based on both styrene and methyl methacrylate (MMA) backbones. When reacted with small molecule cross-linkers, a dynamic network is formed which is both heat-responsive and self-healing. The effect of the polymer backbone and cross-linking density on the formed networks is significant. First generation polymeric FLPs behave in a similar way to the noncovalently linked supramolecular assemblies, where switching from styrene to MMA comonomers creates stiffer polymer networks as the likelihood of chain entanglements increased with a more flexible backbone.



using FLPs as the linking sites.29−33 In our first-generation system, we used polymeric FLPs as macromolecular network precursors with gelation triggered by the addition of a small molecule, diethyl azodicarboxylate (DEAD).29 The FLP crosslinks in our system appeared to be dynamic, producing a material that is temperature responsive and has the capacity to self-heal. Driven by these promising findings, we were motivated to explore the rheological properties of this novel new class of materials. Importantly, we also explored the impact of the nature of the co-monomer on the performance of the poly(FLP) gels. This article represents the first in deep study into the rheological properties of polymeric frustrated Lewis pairs. The macromolecular FLPs in this research were prepared by reversible addition−fragmentation chain transfer (RAFT) copolymerization of designer Lewis base/acid containing monomers with either styrene or methyl methacrylate (MMA) co-monomers. The dynamic behavior of the crosslinks, explored using both rheological and spectroscopic analysis, paints a unique picture of the impressive potential of these first-generation poly(FLP)s.

INTRODUCTION Polymeric materials that can be reshaped, reprocessed, and self-healed after damage have drawn particular attention as a class of smart materials. Such materials contain dynamic crosslinks, which enhance physical properties and respond to external stimuli, enabling reshaping/healing of the material via dynamic exchange of cross-links.1 Compared to conventional, irreversibly cross-linked polymeric networks, dynamic crosslinking recovers the original physical parameters after a damage/healing cycle. There are many dynamic covalent bonds used as cross-links in polymer networks, including carbon−carbon bonds based on reversible Diels−Alder2 or cycloaddition reactions,3 boronic ester 4−6 or boroxine bonding,7 siloxane bonds,8 disulfide/thiol bond.9 In addition, many supramolecular interactions have also been applied into this field of chemistry, including hydrogen bonds,10 ionic bonds,11 π−π stacking,12−14 and Lewis pair complexation.15 Frustrated Lewis pairs (FLPs) were first disclosed by Stephan and co-workers in 2006.16 Compared to conventional Lewis pair adducts, the Lewis acid (LA) and Lewis base (LB) in FLPs bear bulky groups which preclude themselves from binding to each other.16−18 As a result, the LA and LB remain reactive and can activate many small molecules including alkenes,19,20 alkynes,19,21 cyclic ethers,22 carbonyls,18,23 dihydrogen,24,25 and carbon dioxide.26−28 The activated small molecule can bridge between the LA and LB active centers, and in many cases providing dynamic and even reversible linkages under ambient temperatures. We recognized that this unique feature would launch a novel class of dynamic materials © XXXX American Chemical Society



EXPERIMENTAL SECTION

Materials. Styrene (Sigma-Aldrich, ≥99.5%), methyl methacrylate (Sigma-Aldrich, ≥99%), phosphorus trichloride (Sigma-Aldrich, Received: February 7, 2019 Revised: March 29, 2019

A

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

azodicarboxylate (5 equiv) from the stock solution (0.63 M) was quickly injected into the solution. The total concentration of the solution was kept constant at 0.02 moles of FLP units per dm3 after the addition of DEAD solution. The mixture was allowed to stand at ambient temperature overnight. In the event of gel shrinkage, the upper clear solution was decanted off prior to any analysis. Rheology. The mechanical properties of polymeric frustrated Lewis pair gels were analyzed by using a stress-controlled Anton Paar MCR302 rheometer with a 40 mm diameter cross-hatched tool at 1 mm gap. All tests were performed under air, after gels were taken out of the inert atmosphere. The 50 mm diameter bottom plate was likewise cross-hatched, and edges of the gels were covered with inert fluorinated oil (3 M, FC-40) to prevent the evaporation of toluene from the gels. Oscillatory frequency sweeps were performed at 0.1 and 1% strain for 4 and 7% cross-link density gels, respectively, ensuring that both types of gels remained in the linear viscoelastic regime. The frequency range for those experiments was from 0.1 to 100 rad s−1. Amplitude sweep tests were performed at 30 rad s−1, where all of the gels behaved as viscoelastic solids, and the strain was increased from 0.01 to 400%. Recovery tests were performed at 30 rad s−1, with up to 80% strain as this was determined to be enough to yield all of the gels. This was followed by a recovery period where the strain was reduced to 0.1 or 1%, depending on the cross-link density, and the recovery of the moduli was recorded over 5−30 min. Temperature sweeps from 0 to 50 °C were performed at 30 rad s−1 and 0.1 or 1% strain with a heating rate of 0.05 °C s−1.

99%), 4-chlorostyrene (Alfa-Aesar, 99%), 2-bromomesitylene (Acros Organics, 99%), and deuterated chloroform (Sigma-Aldrich, 99.8% D) were dried over CaH2 (Sigma-Aldrich, 95%), then distilled and degassed prior to use. Mg turnings (Sigma-Aldrich, ≥99.5%) were preheated in an oven at 200 °C for at least 2 days before use. Diethyl azodicarboxylate (Alfa-Aesar, 97%) was degassed, then dissolved in anhydrous toluene, and dried with 3 Å molecular sieves. Cumyl dithiobenzoate (CDB), 4-styryl-diphenylborane, 4-styryl-dimesitylphosphine, triphenylborane, and 1,3,5-triisopropyl-2,4,6-trioxane were synthesized according to the literature procedures.29,34−36 Anhydrous diethyl ether, tetrahydrofuran (THF), toluene, and hexane were obtained from an Innovative Technologies Solvent Purification System containing alumina and copper catalysts. All solvents were degassed prior to use by freeze−pump−thaw for three cycles. Anhydrous deuterated toluene was degassed and dried over 3 Å molecular sieves. 2,2′-Azobis(2-methylpropionitrile) (AIBN, SigmaAldrich, 98%) was purified by repeated recrystallization from methanol. 1,2-Dibromoethane (Sigma-Aldrich, 98%), ammonium chloride (Alfa-Aeser, 98%), anhydrous magnesium sulfate (Fisher Scientific), and Fluorinert (3 M, FC-40) used as received. Synthesis of 4-Phenyl-dimesitylphosphine. Under a N2 atmosphere, 1,2-dibromoethane (0.12 mL, 1.39 mmol) was added into a mixture of magnesium turnings (0.40 g, 16.5 mmol) and anhydrous THF (28 mL). The mixture was stirred for 30 min, and then bromobenzene (2.14 g, 13.6 mmol) was added dropwise across 45 min. After addition, the mixture was further refluxed for 30 min, and then stirred at room-temperature (rt) for 1 h to give a gray black Grignard solution. This Grignard solution was transferred dropwise via cannula into a THF solution (40 mL) of dimesitylphosphorus halide (3.60 g, 11.4 mmol) at 0 °C, prepared according to the published procedure.29 After addition, the mixture was allowed to warm back to rt and stirred overnight. The reaction was then quenched by the addition of saturated NH4Cl aqueous solution. The aqueous phase was extracted by Et2O (50 mL × 3), and the combined organic phase was dried over MgSO4. After removal of the solvent under reduced pressure, the crude product was purified by column chromatography (hexane) to give a white solid (1.60 g, 40.6%). Co-polymerization with Styrene. Preweighed amounts of styrene, cumyl dithiobenzoate, 1,3,5-triisopropyl-2,4,6-trioxane (internal standard), and 4-styryl-diphenylborane or 4-styryl-dimesitylphosphine were mixed together under a N2 atmosphere. A small amount of toluene was added to improve the solubility of the Lewis acid/base monomer to a final concentration 7.4 M. A small aliquot of the solution was collected and analyzed by 1H NMR (CDCl3) spectroscopy. The remaining solution was transferred into an ampoule and sealed under an inert atmosphere. The mixture was heated to 110 °C for 45 h. The polymerization was then terminated by cooling the resultant sticky mixture to rt after which a small aliquot was collected for 1H NMR (CDCl3) analysis. The mixture was diluted with a small amount of dry toluene, then precipitated twice into anhydrous hexane under N2 atmosphere. Filtration and subsequent drying under vacuum isolated the co-polymer product as a pink powder. Co-polymerization with MMA. Preweighed amounts of MMA, cumyl dithiobenzoate, 1,3,5-triisopropyl-2,4,6-trioxane (internal standard), AIBN, and 4-styryl-diphenylborane or 4-styryl-dimesitylphosphine were mixed together under N2 atmosphere. A small amount of toluene was added to improve the solubility of the Lewis acid/base monomer to a final concentration 5.7 M. The solution was transferred into an ampoule and sealed under an inert atmosphere. The mixture was heated to 70 °C for 18 h. The polymerization was then terminated by cooling the resultant sticky mixture to rt after which a small aliquot was collected for 1H NMR (CDCl3) analysis. The mixture was diluted with a small amount of toluene, then precipitated twice into hexane under a N2 atmosphere. The copolymer product was obtained as a pink powder by filtration and dried under vacuum. Network Formation. Phosphorus- and boron-containing copolymers were dissolved into anhydrous toluene with 1:1 equiv of boron and phosphorus moieties under a N2 atmosphere. Diethyl



RESULTS AND DISCUSSION To synthesize polymeric frustrated Lewis pairs, styrene-based Lewis acidic and Lewis basic monomers (I and II, Scheme 1) were synthesized, as previously reported by our group.29 These monomers were separately co-polymerized with styrene and MMA via RAFT polymerization37,38 to yield polymers containing sterically crowded Lewis acids, poly(sty-co-LA) and poly(MMA-co-LA), and sterically crowded Lewis bases, poly(sty-co-LB) and poly(MMA-co-LB), as shown in Scheme 1. CDB was selected as the RAFT agent as it enables wellcontrolled polymerizations of styrene and methyl methacrylate, and it was found not to be able to bind with the Lewis acidic borane monomer (I) at both ambient and polymerization temperatures. These first polymeric FLPs reported were based on styrene.29,30 However, as polystyrene is a hard and inflexible polymer, we wanted to contrast these with a classically more flexible backbone, poly(methyl methacrylate) (PMMA).39 We wondered whether the change from styrene to methyl methacrylate would bring some flexibilities to the polymer backbone, as this would enable finer dynamic rearrangement of the polymer chains during the cross-linking process. The copolymerizations of I and II, separately, with styrene were performed at 110 °C. The conditions were slightly modified to 70 °C for the co-polymerizations with MMA since selfinitiation with this co-monomer would result in uncontrolled polymerizations. The radical initiator AIBN was used in these co-polymerizations, as it was found not to affect the gelation as it was not found to be able to bind to the borane monomer in the NMR spectroscopy studies. Different loadings of I and II in co-polymers were targeted to understand the effect of the cross-linking density on the resulting gels. The results of the co-polymerizations are shown on Table 1. Electron-donating or withdrawing substituents on a styrenebased monomer can lead to inductive effects that stabilize the transition states of polymerization propagations.41 Therefore, both monomers I and II likely favor cross-propagation. This is confirmed from the polymerization conversion results shown in Table 1. In general, reasonably low dispersity values were B

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

showed significant disagreements in these values. It is likely that bulky aryl-substituted Lewis acidic and Lewis basic monomers disrupt the organization of the MMA polymer chains in solution, resulting in different hydrodynamic volumes. In addition, GPC analysis was performed by using the dn/dc value known for poly(methyl methacrylate) (0.088 mL mg−1 in THF),40 whereas the observed dn/dc values have a large discrepancy with that of pure PMMA. The discrepancy between theoretical Mn and GPC measured values increases with the borane/phosphine content, supporting this hypothesis. An exemplar dn/dc value was calculated for the 7 mol % phosphine containing co-polymer, changing from 0.088 to 0.103 mL mg−1. Further information about the determination of this value is provided in Supporting Information (SI), Section B. Cross-linking of the poly(FLP) gradient co-polymers was performed using the aforementioned azo compound DEAD as a cross-linker. Phosphorus and boron containing co-polymers were mixed in dry toluene with 1:1 equiv of borane and phosphine moieties. We note that the concentration of the toluene solution after DEAD addition is an important variable to this gelation, and we used a concentration of 0.02 M of FLP moieties (i.e., controlling the concentration of boron and phosphorus atoms in the solution rather than the total polymer) in all gelation experiments. The origin of this cross-linking reaction is the nucleophilic attack of the phosphine monomer to the NN double bond, increasing the nucleophilicity of the second nitrogen (Figure 1A) which is then free to interact with the boron center. Upon injection of DEAD into the co-polymer solution (Figure 1B1,B2) a kinetically frustrated state of the polymer chains is quickly formed. This is followed by the rearrangement of the polymer chains to reach to a more thermodynamically stable state, often leading to gel shrinkage and expulsion of solvent (also see Figure S4). The gel can be reshaped, as shown in Figure 1B4. A series of polymer networks were prepared by using the synthesized macromolecular FLPs, as shown on Table 2. We have prepared 4 and 7 mol % cross-linking poly(sty-co-FLP) networks and 4, 7, and 11 mol % cross-linking poly(MMA-coFLP) networks based on the in-chain ratio of the monomers in each polymer chain. At higher cross-link densities, gels do not form and instead precipitation of a network polymer is favored in both styrene29 and MMA-based FLPs.

Scheme 1. RAFT Co-polymerizations of 4-Styryldiphenylborane, I, and 4-Styryl-dimesitylphosphine, II, with Styrene and MMA, Separatelya

Conditions, (a, c) CDB, 110 °C; (b, d) CDB, AIBN, 70 °C.

a

obtained, suggesting controlled polymerizations. However, the co-polymerizations with low loadings of I and II had higher dispersity values suggesting the possibility of some nonhomogeneities, as some chains would potentially not contain any borane and phosphine monomers. A good agreement between theoretical and experimental number-average molecular weights (Mn) was observed for the polystyrene copolymers. However, MMA-based macromolecular FLPs

Table 1. Polymers Obtained by the RAFT Co-polymerizations of Styrene or MMA with 4-Styryl-diphenylborane or 4-Styryldimesitylphosphinea feeding ratio entry no

polymer

sty/MMA

LA/LB

CDB

1 2 3 4 5 6 7 8 9 10

poly(sty-co-LA) 1 poly(sty-co-LA) 2 poly(sty-co-LB) 1 poly(sty-co-LB) 2 poly(MMA-co-LA) 1 poly(MMA-co-LA) 2 poly(MMA-co-LA) 3 poly(MMA-co-LB) 1 poly(MMA-co-LB) 2 poly(MMA-co-LB) 3

97 95 97 95 98 95 92 97 95 92

3 5 3 5 2 5 8 3 5 8

0.35 0.32 0.35 0.35 0.35 0.35 0.35 0.35 0.35 0.35

conversion (%) AIBN

sty/MMA

LA/LB

Mn (theo) (g mol−1)

Mn (GPC) (g mol−1)

dispersity

0.14 0.14 0.14 0.14 0.14 0.14

65 51 62 60 70 62 57 72 75 76

77 78 87 78 95 98 92 98 96 99

20 800 19 300 21 900 26 900 21 300 21 400 20 900 23 400 25 300 28 700

24 000 20 000 25 000 26 000 29 000 26 000 33 000 37 000 36 000 54 000

1.2 1.2 1.1 1.1 1.3 1.2 1.2 1.6 1.3 1.3

a

Mn (gel permeation chromatography, GPC) values were determined by using triple detector calibration constructed by using very narrow polystyrene standards. The dn/dc values used were 0.185 and 0.088 mL mg−1 for the styrene and MMA-based co-polymers, respectively.40 C

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 1. Cross-linking reaction which results in the formation of polymeric frustrated Lewis pair gels (A) and the visual changes that happen throughout the gelation (B); (1) a mixture of the co-polymers is dissolved in toluene, (2) the addition of DEAD results in an immediate change from the solution to gel, (3) the gel shrinks to a smaller size as a result of the arrangement of the polymer chains arising from the dynamic nature of the cross-linking, and (4) the resulting gel is reshaped and analyzed by a rheometer.

Table 2. Details of the Polymer Networks Prepared by the Addition of a Cross-Linker into the Mixture of Two Co-polymers polymer network poly(sty-co-FLP) 1 poly(sty-co-FLP) 2 poly(MMA-co-FLP) 1 poly(MMA-co-FLP) 2 poly(MMA-co-FLP) 3 poly(sty/MMA-co-FLP)

polymers (entry no) 1 2 5 6 7 4

and and and and and and

cross-linking (mol %)

average number of cross-linking units per polymer

polymer concentration (g cm−3)

Me (g mol−1)

4 7 4 7 11 7

7 12 7 14 22 13

0.12 0.16−0.25

77 000 48 000−31 000

0.14−0.21

6000−10 000

0.12−0.18

6000−9000

3 4 8 9 10 6

G0 ∝ νkBT, where ν is the number density of strands contributing to the elastic network, whereas the relaxation rate depends on the time scale for bond dissociation.42,46 We repeated small amplitude frequency sweeps at different temperatures from 0 to 50 °C (SI Figures S6−S10). Above 50 °C, even with the oil seal, toluene would rapidly evaporate and dry out the gel. Although the frequency-dependent moduli change with temperature in absolute terms, curves at different temperatures can all be collapsed onto a master curve using time−temperature superposition (TTS), (Figures 2B and S11).46 This collapse is achieved by rescaling the absolute frequencies, ω → aTω, where aT is a temperature-dependent shift factor relative to a reference temperature. The shift factor decreases with increasing temperature, and a plot of log aT vs temperature is provided in Figure S12. As the poly(FLP) gels obey TTS, it can be inferred that they do not undergo any phase transitions or significant structural changes over this temperature range (0−50 °C). The master curve can again be fitted with the single-mode Maxwell model, and the frequency shifts reflect a decrease in the relaxation time scale, tc(T) = 1/ ωc(T), with increasing temperature. This is frequently observed in supramolecular gels formed from solutions of associating polymers,47 where the network contains a number of noncovalent interactions with finite lifetimes.48 We further probed this temperature dependence, recording G′(T) and G″(T) at a fixed frequency, ω = 30 rad s−1, as the temperature was slowly increased from 0 to 50 °C (Figure 2C). The results are consistent with our TTS shifted master curve,

Dynamic mechanical properties of polymeric frustrated Lewis pair networks were investigated using oscillatory rheology. Results are shown for the 4 mol % cross-linking poly(sty-co-FLP) network in Figure 2, with similar behaviors observed in other poly(FLP) networks explored here (see Figures S5 and 4D for more examples). Under small amplitude oscillatory shear, the frequency-dependent storage and loss moduli, G′(ω) and G″(ω), characterize the elastic (solid-like) and viscous (liquid-like) material responses. These are shown for 4 mol % cross-linking poly(sty-co-FLP) in Figure 2A. At high frequencies, the storage modulus dominates, G′ > G″, implying that most of the deformation energies are stored elastically over short time scales, whereas at low frequencies the loss modulus dominates, G″ > G′, implying that most of the energies are dissipated through the relaxation of the polymer chains over longer time scales. The cross-over frequency, ωc, is defined by the condition G′(ωc) = G″(ωc), and for this particular poly(FLP) network, ωc = 19 rad s−1 at 20 °C. The moduli can be fitted by the single-mode Maxwell model G′(ω) = G0(ω/ωc)2 /(1 + (ω/ωc)2 )

G″(ω) = (ω/ωc)G′

with G0 = 3800 Pa.42,43 This single-mode behavior is typically observed in wormlike micelles and also has been observed in polymeric supramolecular assemblies, such as hydrogenbonded polymers.44,45 The plateau modulus should scale as D

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 2. Dynamic mechanical properties of 4% cross-linking poly(sty-co-FLP) obtained by oscillatory rheology. Frequency dependencies of the dynamic moduli of the polymer network at 20 °C, 1% strain and theoretical fit obtained by using the Maxwell model (A), ωc = 19 rad s−1, G0 = 3800 Pa, time−temperature superposition (TTS) at the reference temperature 20 °C (B), temperature dependencies of the dynamic moduli G′ and G″ at 30 rad s−1, 1% strain at a heating rate of 0.05 °C s−1 (C), oscillatory amplitude sweep test showing how dynamic moduli and shear stress behave as a function of strain, (D) and the recovery of G′ and G″ after failure induced by shear strain (E). The polymeric gel was sheared to failure under increasing oscillatory strain followed by a recovery under 1% strain.

with both moduli decreasing as the temperature increases, and ultimately crossing at 32 °C. Using the fact that the storage and loss moduli in a single mode Maxwell model are simply related by G′ = ωtcG″, we estimate the temperature dependence of the relaxation time tc(T) = G′(T)/ωG″(T). In general, relaxation of the polymer network involves contributions from segmental motion of the long polymer backbones, reptation, (tr) and dissociation (td) times of the cross-links.49,50 If the time scale for breaking cross-links is sufficiently fast (td ≪ tr), the relaxation dynamics can be characterized by a single relaxation time corresponding to a first-order or pseudo-first-order reaction rate, kobs = 1/tc.42 Fitting our results to an Arrhenius equation kobs(T) = k0 e−Ea/RT, it is possible to relate kobs to the activation energy (Ea) for this process (Figure S13), calculated as 18.2 kJ mol−1. This relaxation time, and thus also Ea, characterize the combined time scales for reptation and dissociation. This low value of Ea is realistic for a reversible process.42 The significant effect of the temperature suggests that the interaction between the anion and boron center is truly dynamic. An NMR spectroscopy experiment was designed to observe this interaction. Model compounds 4-phenyl-dimesitylphosphine and triphenylborane were synthesized to mimic the interactions of polymeric FLPs. These molecules were mixed with 1 equiv of DEAD, and the interaction between the borane center and the nitrogen anion monitored at different temperatures by NMR spectroscopy (Figure 3). As temperature increases, the diagnostic resonance (ca. 1 ppm) decreases in intensity, accompanied by the formation of free triphenylborane (ca. 67 ppm) at high temperatures. These results confirm the temperature dependency of the interaction and its highly dynamic nature.

Figure 3. 11B NMR spectra to demonstrate the effect of temperature on the cross-linking. Model compounds, triphenylborane and 4phenyl-dimesitylphosphonine, were synthesized and linked by using 1 equiv of DEAD in anhydrous toluene-d8. Variable temperature NMR experiments were performed from 27 to 97 °C.

Performing an amplitude sweep at 30 rad s−1, where the polymer network behaves as a viscoelastic solid, we find that G′ and G″ remained constant up to a yield strain around 28%. Beyond this linear viscoelastic region, the material suddenly yields, with a rapid drop in the moduli and a decrease in the applied stress. This sudden stress drop, occurring at a yield stress of 1050 Pa, suggests the sudden fracture of the network, possibly due to degradation of non-cross-linked chains or E

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 4. Plots of changes in cross-link density and polymer backbone vs. the dynamic mechanical properties of the polymeric frustrated Lewis pair gels. Frequency (A, D) and amplitude (B, C, E, F) sweep tests are shown above to describe these systems. Plots (A−C) show the effect of changing the cross-link density from 4 to 7 mol % on styrene-based polymeric FLP gel, whereas the plots (D−F) show the effect of polymer backbone where three different gels were prepared and reported, one based on PMMA, one based on polystyrene, and other a mixture of the two polymer backbones, all with 7 mol % cross-linking.

increase in the reptation time as the polymer network becomes increasingly restrictive. Comparing strain amplitude sweeps with the two gels (Figure 4B,C), we find a higher yield stress in the higher cross-link density gel but a similar yield strain, γy ∼ 20%. From the plateau modulus and the polymer concentration, c, we can estimate the molecular weight between effective cRT entanglements, Me, using G0 = M (Table 2). In gels with

dissociation of noncovalent bonds. Relatively low yield strains, here γy ∼ 28%, have been reported in other supramolecular polymer gels.48 In our previous report, we showed that polymeric frustrated Lewis pair gels can self-heal.29 In this study, we further probe the self-healing properties of our polymeric system. We performed a series of shear amplitude sweeps, increasing the strain from 1 to 80% over 160 s to yield the gels, then recording the recovery of the moduli with time at a fixed 1% amplitude.51 Results for a 4 mol % cross-linking poly(sty-coFLP) are shown in Figure 2E. As the shear amplitude is increased beyond the yield strain, G′ and G″ decreased significantly, but after the strain returns to the linear viscoelastic region, network reconstitution began instantaneously. The gel recovers nearly all of its former strength roughly 15 min after each yielding step. On average, 89 and 91% recoveries were observed after each yielding step for G′ and G″, respectively, demonstrating repeatable self-healing in these poly(FLP) gels after damage. Having synthesized a variety of co-polymers, we explore how a change in cross-link density and polymer backbone affect the mechanical properties of our supramolecular polymeric gels. To examine the role of the cross-link density, poly(sty-co-FLP) gels with 4 and 7 mol % cross-link densities were prepared by using the pairs 1, 3 and 2, 4 from Table 1. Comparing frequency sweeps at 20 °C for the two gels (Figure 4A), we find that the cross-over frequency shifts to a lower frequency with an increase in cross-linking density, along with a concomitant increase in the magnitude of the moduli. The decrease in ωc with cross-linker density, corresponding to an increase in relaxation time tc, has been previously observed in polymeric supramolecular assemblies.43 This increase in relaxation time could reflect either the need for multiple simultaneous bond breakages to effectively free chains or an

e

higher cross-linking densities, phase separation and gel shrinkage make it challenging to accurately determine the polymer concentration c. This is both due to the variable polymer concentration in the discarded solvent and also solvent-rich regions that tended to remain in the vial after loading the gel into the rheometer. In these gels, a range of concentration and Me values was provided, reflecting this uncertainty. The values obtained here, 77 000 and ∼40 000 g mol−1 for 4 and 7 mol % cross-link densities, are larger than the molecular weights of the individual polymer chains, indicating a relatively low density of effective cross-links. Although there are multiple LA/LB sites per chain, the average number of these units are 7 and 12 for the two network systems. Each polymer chain contains over 180 units and with this sparse presence of the LA/LB units, it is reasonable to expect large Me values. The decrease in Me with increasing cross-linker concentration reflects an increase in the density of network junctions with an increasing number of reversible FLP bonds between chains. We have demonstrated that these poly(FLP) networks exhibit numerous rheological features found in supramolecular physical gels formed by associating polymers in solution, reflecting the role of the reversible, non-covalent FLP interactions. The low density of effective cross-links as pointed out in Table 2, along with the approximate agreement with a F

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

generally observed in supramolecular physical gels and are generally ascribed to new relaxation modes beginning to enter the dynamics.42,47 Deviations at high frequencies are also present in even the 4% cross-linker styrene-backbone gels (Figure 4A) but are much more apparent in these MMAbackbone gels. In these gels, it is likely that the higher density of cross-links likely begins to complicate the network relaxation dynamics. Following the yielding behavior of the 7 mol % polymer networks, we wondered how a change in cross-link density and backbone would affect the self-healing properties of the gels. Recovery results similar to the 4 mol % poly(sty-co-FLP) network were obtained. As shown in Figure 6, the gels quickly

single-mode Maxwell model, indicates that these networks can be reasonably well-described by either the transient network model52 or worm-like micelle inspired models.44,53 This is in contrast to behavior typically observed in associative polymer melts,50 which are instead frequently modeled with a ‘sticky Rouse’ model describing the hindered reptation of polymers with multiple binding sites per chain.54,55 To investigate the role of the polymer backbone, gels were prepared at 7 mol % cross-link density with the following pairs of co-polymers 2, 4 (both styrene), 6, 9 (both MMA), and 6, 4 (mixed styrene-MMA). Theoretical molecular weights of the co-polymers are similar, suggesting that any behavioral differences would be dominated by the change of backbone. Frequency sweeps were run on all resultant gels at 20 °C (Figure 4D). The cross-over frequency was the lowest in the poly(MMA-co-FLP) gel and the highest in the poly(sty-coFLP) gel, with the mixed poly(sty/MMA-co-FLP) gel falling in between, indicating a longer relaxation time with the more flexible MMA backbone. The gels with the MMA backbone exhibited a higher plateau modulus G0, translating to a significantly lower molecular weight between effective entanglements, with Me dropping from ∼40 000 g mol−1 with the styrene backbone to ∼8000 g mol−1 with the MMA backbone. This indicates that the more flexible backbone actually results in a more solid-like gel by allowing the formation of more cross-links. To investigate the yielding behavior with different backbones, we carried out strain sweeps on these different gels (Figure 4D,E). We found that networks formed with MMA containing co-polymers resulted in more brittle gel structures, as poly(MMA-co-FLP), poly(sty/MMA-co-FLP), and poly(styco-FLP) gels yielded at 1, 22, and 28% strains, respectively. The yield stresses of both the MMA and styrene backbone gels were comparable, ∼2 kPa, but with the mixed backbone the yield stress increased to ∼10 kPa. Carrying out small amplitude frequency sweeps with the mixed backbone poly(sty/MMA-co-FLP) 7 mol % crosslinking networks, we find that these mixed backbone gels also obey the time−temperature superposition (Figures 5 and S14−S18). However, at high frequencies, there are notable deviations from the single-mode Maxwell behavior, with G″ flattening to a plateau instead of decaying as ω−1. Deviations from simple single-mode dynamics at high frequencies are

Figure 6. Self-healing properties of the gels with 7 mol % cross-linking (A) poly(sty-co-FLP), (B) poly(sty/MMA-co-FLP), and (C) poly(MMA-co-FLP) polymer networks. The gels were exposed to a stain value of 400% and the recovery of G′ and G″ was recorded after the failure. The experiment was performed at 30 rad s−1 over 1200 s.

healed themselves upon the removal of strain. Recoveries (91, 90, and 86%) were observed for the 7 mol % cross-link density poly(sty-co-FLP), poly(sty/MMA-co-FLP), and poly(MMA-coFLP) polymer networks, respectively. These values were obtained in 15 min, proving the fast healing characteristic of the poly(FLP) gels. Notably, there remained a positive healing trend when the experiment was terminated, suggesting potential full recovery over longer healing times. However, as toluene in the gels can evaporate, this could also cause an increase in G′ and G″ values and thus we purposely kept healing experiments short to mitigate this influence. The results are in line with the previously reported self-healing results with the networks formed by reversible cross-links.51,56 We note that any decrease in the gel stiffness could arise from

Figure 5. Time−temperature superposition master plot of frequency dependencies of the dynamic moduli of the 7 mol % cross-linking mixed sty/MMA-based polymeric FLP network at a reference temperature of 20 °C. G

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Notes

the hydrolysis of the borane units in the polymer networks. All rheology experiments were performed in air with a humidity up to 46%, giving potential for hydrolysis of the borane units within the polymer networks. These borane units are susceptible to water in the air as once they are not protected by DEAD, i.e., during rearrangement/shearing. It is possible that every time some borane units are lost, and the number of cross-links decreases.

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We kindly thank the Universities of Edinburgh and Manchester for financial support.





CONCLUSIONS A variety of co-polymers with either sterically encumbered Lewis acid or Lewis base pendant functional groups were synthesized and activated using a small molecule, DEAD, resulting in a dynamically cross-linked network structure. Rheological behavior of the resulting polymer networks showed that the networks behave as noncovalently linked supramolecular assemblies with a high-temperature dependence. The cross-linking was shown to be highly dynamic. The self-healing behaviors of the gels were characterized, and the healing is fast and happens almost immediately after the removal of yielding strain. The network structure breaks at relatively low strains, a characteristic of supramolecular assemblies. Poly(FLP) gels were understood to obey the time−temperature superposition suggesting that no change happens in the microstructure of the polymer network. It was demonstrated that an increase in cross-link density resulted in an increase in the likelihood of entanglements, resulting in a more solid-like material. A switch in polymer backbone from inflexible styrene to relatively more flexible MMA units results in easier chain rearrangement and cross-link exchange, where the material was almost solid-like at all frequencies analyzed. We believe that the combination of rich chemistry of FLPs with known activations of small molecules, such as H2, CO2, N2O, alkenes and imines, and wide range of synthetic materials available through altering macromolecular backbone and architecture, such as blocks, stars and surfaces, promises a whole new class of novel stimuli-responsive materials with potential applications including CO2 capture and catalysis.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.9b00271.



REFERENCES

(1) Wojtecki, R. J.; Meador, M. A.; Rowan, S. J. Using the dynamic bond to access macroscopically responsive structurally dynamic polymers. Nat. Mater. 2010, 10, No. 14. (2) Chen, X.; Dam, M. A.; Ono, K.; Mal, A.; Shen, H.; Nutt, S. R.; Sheran, K.; Wudl, F. A Thermally Re-mendable Cross-Linked Polymeric Material. Science 2002, 295, 1698−1702. (3) Chung, C.-M.; Roh, Y.-S.; Cho, S.-Y.; Kim, J.-G. Crack Healing in Polymeric Materials via Photochemical [2+2] Cycloaddition. Chem. Mater. 2004, 16, 3982−3984. (4) Röttger, M.; Domenech, T.; van der Weegen, R.; Breuillac, A.; Nicolaÿ, R.; Leibler, L. High-performance vitrimers from commodity thermoplastics through dioxaborolane metathesis. Science 2017, 356, 62−65. (5) Cash, J. J.; Kubo, T.; Bapat, A. P.; Sumerlin, B. S. RoomTemperature Self-Healing Polymers Based on Dynamic-Covalent Boronic Esters. Macromolecules 2015, 48, 2098−2106. (6) Chen, W.-P.; Hao, D.-Z.; Hao, W.-J.; Guo, X.-L.; Jiang, L. Hydrogel with Ultrafast Self-Healing Property Both in Air and Underwater. ACS Appl. Mater. Interfaces 2018, 10, 1258−1265. (7) Ogden, W. A.; Guan, Z. Recyclable, Strong, and Highly Malleable Thermosets Based on Boroxine Networks. J. Am. Chem. Soc. 2018, 140, 6217−6220. (8) Zheng, P.; McCarthy, T. J. A Surprise from 1954: Siloxane Equilibration Is a Simple, Robust, and Obvious Polymer Self-Healing Mechanism. J. Am. Chem. Soc. 2012, 134, 2024−2027. (9) Canadell, J.; Goossens, H.; Klumperman, B. Self-Healing Materials Based on Disulfide Links. Macromolecules 2011, 44, 2536−2541. (10) Sijbesma, R. P.; Beijer, F. H.; Brunsveld, L.; Folmer, B. J. B.; Hirschberg, J. H. K. K.; Lange, R. F. M.; Lowe, J. K. L.; Meijer, E. W. Reversible Polymers Formed from Self-Complementary Monomers Using Quadruple Hydrogen Bonding. Science 1997, 278, 1601−1604. (11) Wei, Z.; He, J.; Liang, T.; Oh, H.; Athas, J.; Tong, Z.; Wang, C.; Nie, Z. Autonomous self-healing of poly(acrylic acid) hydrogels induced by the migration of ferric ions. Polym. Chem. 2013, 4, 4601− 4605. (12) Greenland, B. W.; Burattini, S.; Hayes, W.; Colquhoun, H. M. Design, synthesis and computational modelling of aromatic tweezermolecules as models for chain-folding polymer blends. Tetrahedron 2008, 64, 8346−8354. (13) Burattini, S.; Greenland, B. W.; Merino, D. H.; Weng, W.; Seppala, J.; Colquhoun, H. M.; Hayes, W.; Mackay, M. E.; Hamley, I. W.; Rowan, S. J. A Healable Supramolecular Polymer Blend Based on Aromatic π−π Stacking and Hydrogen-Bonding Interactions. J. Am. Chem. Soc. 2010, 132, 12051−12058. (14) Fox, J.; Wie, J. J.; Greenland, B. W.; Burattini, S.; Hayes, W.; Colquhoun, H. M.; Mackay, M. E.; Rowan, S. J. High-Strength, Healable, Supramolecular Polymer Nanocomposites. J. Am. Chem. Soc. 2012, 134, 5362−5368. (15) Vidal, F.; Lin, H.; Morales, C.; Jäkle, F. Polysiloxane/ Polystyrene Thermo-Responsive and Self-Healing Polymer Network via Lewis Acid-Lewis Base Pair Formation. Molecules 2018, 23, No. 405. (16) Welch, G. C.; Juan, R. R. S.; Masuda, J. D.; Stephan, D. W. Reversible, Metal-Free Hydrogen Activation. Science 2006, 314, 1124−1126. (17) Stephan, D. W. The broadening reach of frustrated Lewis pair chemistry. Science 2016, 354, No. aaf7229.

Additional analytical characterization (Section A); determination of dn/dc value for the poly(MMA-coLB) (Section B); data for the shrinkage of the 7 mol % cross-linking poly(MMA-co-FLP) network (Section C); supplementary rheology data and analysis (Section D) (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Meng Wang: 0000-0002-5926-4970 Michael P. Shaver: 0000-0002-7152-6750 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. H

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (18) Stephan, D. W.; Erker, G. Frustrated Lewis Pair Chemistry: Development and Perspectives. Angew. Chem., Int. Ed. 2015, 54, 6400−6441. (19) Mömming, C. M.; Frömel, S.; Kehr, G.; Fröhlich, R.; Grimme, S.; Erker, G. Reactions of an Intramolecular Frustrated Lewis Pair with Unsaturated Substrates: Evidence for a Concerted Olefin Addition Reaction. J. Am. Chem. Soc. 2009, 131, 12280−12289. (20) McCahill, J. S. J.; Welch, G. C.; Stephan, D. W. Reactivity of “Frustrated Lewis Pairs”: Three-Component Reactions of Phosphines, a Borane, and Olefins. Angew. Chem., Int. Ed. 2007, 46, 4968−4971. (21) Dureen, M. A.; Stephan, D. W. Terminal Alkyne Activation by Frustrated and Classical Lewis Acid/Phosphine Pairs. J. Am. Chem. Soc. 2009, 131, 8396−8397. (22) Birkmann, B.; Voss, T.; Geier, S. J.; Ullrich, M.; Kehr, G.; Erker, G.; Stephan, D. W. Frustrated Lewis Pairs and Ring-Opening of THF, Dioxane, and Thioxane. Organometallics 2010, 29, 5310−5319. (23) Moebs-Sanchez, S.; Bouhadir, G.; Saffon, N.; Maron, L.; Bourissou, D. Tracking reactive intermediates in phosphine-promoted reactions with ambiphilic phosphino-boranes. Chem. Commun. 2008, 3435−3437. (24) Lam, J.; Szkop, K. M.; Mosaferi, E.; Stephan, D. W. FLP catalysis: main group hydrogenations of organic unsaturated substrates. Chem. Soc. Rev. 2019, No. 277. (25) Stephan, D. W.; Erker, G. Frustrated Lewis Pairs: Metal-free Hydrogen Activation and More. Angew. Chem., Int. Ed. 2010, 49, 46− 76. (26) Ashley, A. E.; O’Hare, D. In FLP-Mediatied Activations and Reductions of CO2 and CO, 1st ed.; Stephan, D. W., Ed.; Springer: Berlin, 2013; pp 191−217. (27) Mömming, C. M.; Otten, E.; Kehr, G.; Fröhlich, R.; Grimme, S.; Stephan, D. W.; Erker, G. Reversible Metal-Free Carbon Dioxide Binding by Frustrated Lewis Pairs. Angew. Chem., Int. Ed. 2009, 48, 6643−6646. (28) Stephan, D. W.; Erker, G. Frustrated Lewis pair chemistry of carbon, nitrogen and sulfur oxides. Chem. Sci. 2014, 5, 2625−2641. (29) Wang, M.; Nudelman, F.; Matthes, R. R.; Shaver, M. P. Frustrated Lewis Pair Polymers as Responsive Self-Healing Gels. J. Am. Chem. Soc. 2017, 139, 14232−14236. (30) Chen, L.; Liu, R.; Yan, Q. Polymer Meets Frustrated Lewis Pair: Second-Generation CO2-Responsive Nanosystem for Sustainable CO2 Conversion. Angew. Chem., Int. Ed. 2018, 57, 9336−9340. (31) Chen, L.; Liu, R.; Hao, X.; Yan, Q. CO2-Cross-Linked Frustrated Lewis Networks as Gas-Regulated Dynamic Covalent Materials. Angew. Chem., Int. Ed. 2018, 13, 270−274. (32) Liu, R.; Liu, X.; Ouyang, K.; Yan, Q. Catalyst-Free Click Polymerization of CO2 and Lewis Monomers for Recyclable C1 Fixation and Release. ACS Macro Lett. 2019, 8, 200−204. (33) Niu, Z.; Bhagya Gunatilleke, W. D. C.; Sun, Q.; Lan, P. C.; Perman, J.; Ma, J.-G.; Cheng, Y.; Aguila, B.; Ma, S. Metal-Organic Framework Anchored with a Lewis Pair as a New Paradigm for Catalysis. Chem 2018, 4, 2587−2599. (34) Saindane, P.; Jagtap, R. N. RAFT copolymerization of amphiphilic poly (ethyl acrylate-b-acrylic acid) as wetting and dispersing agents for water borne coating. Prog. Org. Coat. 2015, 79, 106−114. (35) Arias-Ugarte, R.; Wekesa, F. S.; Findlater, M. Selective aldol condensation or cyclotrimerization reactions catalyzed by FeCl3. Tetrahedron Lett. 2015, 56, 2406−2411. (36) Brown, H. C.; Racherla, U. S. Organoboranes. 44. A convenient, highly efficient synthesis of triorganylboranes via a modified organometallic route. J. Org. Chem. 1986, 51, 427−432. (37) Chiefari, J.; Chong, Y. K.; Ercole, F.; Krstina, J.; Jeffery, J.; Le, T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.; Moad, G.; Rizzardo, E.; Thang, S. H. Living Free-Radical Polymerization by Reversible Addition−Fragmentation Chain Transfer: The RAFT Process. Macromolecules 1998, 31, 5559−5562. (38) Cheng, F.; Bonder, E. M.; Jäkle, F. Electron-Deficient Triarylborane Block Copolymers: Synthesis by Controlled Free

Radical Polymerization and Application in the Detection of Fluoride Ions. J. Am. Chem. Soc. 2013, 135, 17286−17289. (39) Hasan, O. A.; Boyce, M. C. Energy storage during inelastic deformation of glassy polymers. Polymer 1993, 34, 5085−5092. (40) Coward, D. L.; Lake, B. R. M.; Shaver, M. P. Understanding Organometallic-Mediated Radical Polymerization with an Iron(II) Amine−Bis(phenolate). Organometallics 2017, 36, 3322−3328. (41) Brandrup, J.; Immergut, E. H. In Polymer Handbook; Wiley: NY, 1989. (42) Vermonden, T.; van Steenbergen, M. J.; Besseling, N. A. M.; Marcelis, A. T. M.; Hennink, W. E.; Sudhölter, E. J. R.; Cohen Stuart, M. A. Linear Rheology of Water-Soluble Reversible Neodymium(III) Coordination Polymers. J. Am. Chem. Soc. 2004, 126, 15802−15808. (43) Yount, W. C.; Loveless, D. M.; Craig, S. L. Small-Molecule Dynamics and Mechanisms Underlying the Macroscopic Mechanical Properties of Coordinatively Cross-Linked Polymer Networks. J. Am. Chem. Soc. 2005, 127, 14488−14496. (44) Cates, M. E.; Candau, S. J. Statics and dynamics of worm-like surfactant micelles. J. Phys.: Condens. Matter 1990, 2, No. 6869. (45) Lortie, F.; Boileau, S.; Bouteiller, L.; Chassenieux, C.; Demé, B.; Ducouret, G.; Jalabert, M.; Lauprêtre, F.; Terech, P. Structural and Rheological Study of a Bis-urea Based Reversible Polymer in an Apolar Solvent. Langmuir 2002, 18, 7218−7222. (46) Larson, R. G. In The Structure and Rheology of Complex Fluids; Oxford University Press: NY, 1999. (47) Noro, A.; Matsushita, Y.; Lodge, T. P. Thermoreversible Supramacromolecular Ion Gels via Hydrogen Bonding. Macromolecules 2008, 41, 5839−5844. (48) Noro, A.; Hayashi, M.; Matsushita, Y. Design and properties of supramolecular polymer gels. Soft Matter 2012, 8, 6416−6429. (49) Wu, S.; Liu, S.; Zhang, Z.; Chen, Q. Dynamics of Telechelic Ionomers with Distribution of Number of Ionic Stickers at Chain Ends. Macromolecules 2019, 52, 2265−2276. (50) Zhang, Z.; Chen, Q.; Colby, R. H. Dynamics of associative polymers. Soft Matter 2018, 14, 2961−2977. (51) Holten-Andersen, N.; Harrington, M. J.; Birkedal, H.; Lee, B. P.; Messersmith, P. B.; Lee, K. Y. C.; Waite, J. H. pH-induced metalligand cross-links inspired by mussel yield self-healing polymer networks with near-covalent elastic moduli. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 2651−2655. (52) Tanaka, F.; Edwards, S. F. Viscoelastic properties of physically cross-linked networks: Part 2. Dynamic mechanical moduli. J. NonNewtonian Fluid Mech. 1992, 43, 273−288. (53) Cates, M. E.; Witten, T. A. Chain conformation and solubility of associating polymers. Macromolecules 1986, 19, 732−739. (54) Rubinstein, M.; Semenov, A. N. Dynamics of Entangled Solutions of Associating Polymers. Macromolecules 2001, 34, 1058− 1068. (55) Chen, Q.; Tudryn, G. J.; Colby, R. H. Ionomer dynamics and the sticky Rouse model. J. Rheol. 2013, 57, 1441−1462. (56) Terech, P.; Yan, M.; Maréchal, M.; Royal, G.; Galvez, J.; Velu, S. K. P. Characterization of strain recovery and “self-healing” in a selfassembled metallo-gel. Phys. Chem. Chem. Phys. 2013, 15, 7338−7344.

I

DOI: 10.1021/acs.macromol.9b00271 Macromolecules XXXX, XXX, XXX−XXX